Professional Documents
Culture Documents
Policies of Planing
Policies of Planing
Policies of Planing
2006/33
Annelies Zoomers*
March 2006
Abstract
This article aims to contribute to the discussion about how to make development
interventions more effective by analyzing the factors contributing to the success or
failure of rural development projects. We made an aggregate level analysis of 46
projects in the field of agricultural research (AR), water management (WM), natural
resource management (NRM), and integrated rural development (IRD), financed by the
Netherlands’ Directorate-General for International Cooperation (DGIS) and carried out
between 1975-2005 in Asia, Africa and Latin America. Making a distinction between
the successful projects and failures, we showed the possibilities and limitations …/
Acknowledgements
The views expressed in this publication are those of the author(s). Publication does not imply
endorsement by the Institute or the United Nations University, nor by the programme/project sponsors, of
any of the views expressed.
1 Introduction: the changing policy context
For many years, rural development has been one of the priority goals of Dutch
development co-operation. Rural development projects were seen as important drivers
of development, as they provided the inputs for industrial development and increased
export earnings, and contributed to food security. Stagnating agricultural production,
environmental degradation, and the concentration of poverty in the countryside were
considered obstacles to further national development. Consequently, large investments
were made in agricultural research and extension services, water management, natural
resource management, and integrated rural development.
Since the mid 1990s, it was increasingly acknowledged that isolated projects would not
result in sustainable results, unless they were embedded in a sound macroeconomic
situation and a supportive policy environment (Schulpen 2001; DGIS 2003a). The often
mentioned negative consequences of the project based approach include the patchwork
management of development assistance, inadequate local ownership, the overloading of
local capacity to co-ordinate donor relationships, and the lack of sustainability and
institutional development, all of which result in a waste of development resources
(Euforic 2004; Grinspun 2001; Foster et al. 2000: 31; Mayhew 2002). Against a
backdrop of increasing aid fatigue and heightened concern about development
performance and results, aid agencies began to move resources from project funding to
sector-wide approaches (SWAps), which required donors to pool their funds and make
development intervention part of the receiving countries’ ‘normal’ government policy.
Instead of carrying out development projects, donors nowadays come together to pool
their funds rather than supporting separate programmes. Donors and recipient
governments jointly agree on targets and strategies for allocating the pooled funds and
implementing defined priority projects (DGIS 2003b).
An example of a donor that moved from project funding to the SWAps is the
Netherlands’ Directorate-General for International Co-operation (DGIS). The DGIS
forms part of the Dutch foreign ministry and is responsible for development co-
operation policy which is based on an annual budget of 0.8 per cent of the Netherlands’
GNP (in line with internationally agreed standards). The DGIS decided to introduce the
sectoral approach in 1999, ‘[t]o boost the effectiveness and sustainability of Dutch aid’.
This has meant reducing both the number of countries receiving aid and the number of
beneficiary sectors within recipient countries. This sectoral approach is not an end in
itself but a process by which sector-based assistance can be lent more effectively. ‘It is a
way of integrating aid into the sectoral policies of recipient countries’ (Minbuza 2000: 3).
The Netherlands has selected 36 partner countries on the basis of ‘good governance’
and provided budget support and assistance to sector-wide programmes. The DGIS is
giving emphasis to working with other donors to devise a more cohesive aid package
1
and ultimately to move towards sectoral budget aid. These efforts are largely guided by
the key policy aims of poverty reduction, gender equality and women’s empowerment,
environmental protection, good governance and institutional development, which are
collectively expressed by the Dutch acronym ‘GAVIM’. Priority is given to policy
fields that contribute to the achievement of the Millennium Development goals, with
half of the budget going to Africa (DGIS 2000: 3).
While making attempts to implement the SWAps, the DGIS decided to carry out a field
evaluation of rural development projects realized in the period 1975-2005, in order to
learn from earlier experiences and see how the lessons could be used to improve the
implementation of the new programme-oriented strategy. They took a representative
sample of 46 projects in the field of water management, natural resource management,
agricultural extension and integrated rural development,1 and asked separate teams to
use a similar format in making an assessment of the results, and identify the underlying
success and failure factors.2 Successes and malfunctions were expressed in terms of
relevance, efficiency, effectiveness, sustainability and impact of project interventions,3
while also taking into account the orientation of projects on poverty, gender,
institutional development, governance and environment (the GAVIM-goals, prioritized
by DGIS).
This study is an attempt to learn from experience, i.e., making an aggregate level
assessment of the factors contributing to the success or failure of development projects
in contributing to sustainable development. Insofar as lessons are drawn from
development projects, emphasis is often put on the reasons for failure: Projects will not
contribute to sustainable results, due to the limited scale of intervention (islands of
wealth), the lack of ownership, and the lack of connection between the micro and macro
level (Foster 2000: 41; Mayhew 2002). Little is known about why similarly designed
projects, with similar aims, produce differential results in different settings. Making a
distinction between the successful projects and failures on the basis of DAC and
GAVIM performance, what types of factors explain the different results; what can we
learn to make future policy more effective?
1 This selection was made by the DGIS staff, based on project goals, location, duration, and budget size.
The selected projects represent almost 50 per cent of the total budget spent on agricultural research and
extension services, water management, natural resource management, and integrated rural development
(1975-2005), and covers about 25 per cent of the total number of projects. The sample is restricted to
project support (i.e., not including programme and/or budget support). No further details are available for
reasons of confidentiality, but the sample contains the more important projects in terms of budget and
duration. The representativeness issue did not consider, however, when projects were instigated (1975-
2005), and there is some over-representation of the more recent projects.
2 This article is based on these 46 reports. The author participated in a consultancy assigned to analyse
patterns explaining the success and failure of projects; see for the final reports, van Dijk et al. (2005).
3 For instance, the official Development Assistance Committee (DAC) evaluation criteria.
2
• To facilitate an aggregate-level assessment of the factors contributing to the
success or failure of the projects, we compiled a database on the basis of the 46
evaluation reports containing the following information.
• Background to the context of the projects at country level and, if possible, at the
level of the intervention area (HDI, life expectancy, literacy rate, GDP, macro-
structural reform, political situation, ecological problems, and climate).
• Basic project characteristics, namely duration, budget, type of activities, mission
or goal, donor collaboration, etc.
• Project results (as formulated in the evaluation reports), namely the indicators
used, and the assessment of critical success and failure factors as mentioned in the
report; also the scores on the GAVIM goals and the DAC indicators.
This study first presents a summary of the characteristics of the 46 projects under
review, followed by a description of the DAC and GAVIM results. An attempt is then
made to establish patterns of factors that contribute to the success or failure of
development projects. Finally, it identifies a number of hidden problems which are often
neglected in policy debates. We will show that DAC and GAVIM criteria help to
understand what type of factors play a role in the success or failure of projects; the
outcome will finally depend on more general factors such as how projects are
strategically linked in broader policies, their local embeddedness (i.e., how to respond to
or anticipate complexities), and whether they are in line with the short- and long-term
strategies of the people who are supposed to benefit from the project interventions.
These dimensions are often not sufficiently taken into account in current discussions
about learning in development co-operation.
Since the 1980s, DGIS has been actively trying rural development by implementing
development projects in various fields, notably agricultural research (AR), water
management (WM), natural resource management (NRM), and integrated rural
development (IRD). The DGIS has implemented these projects in a large number of
countries which were usually confronted by adverse economic, political, and
environmental circumstances (Table 1). Using the HDI ranking as an indicator for the
socioeconomic situation, we see that many projects were carried out in the poorest
countries (Mozambique, Mali, Burkina Faso, Ethiopia), although there were also
interventions in relatively richer countries (Ecuador, Philippines, Costa Rica). Most
countries were in a post-conflict or even war situation (Ethiopia, Mozambique,
Nicaragua, El Salvador); experienced a lot of political turbulence during the period of
project implementation (Pakistan, Philippines, Mali, Kenya, Nepal, India) and/or had to
deal with adverse climatic/environmental conditions, namely a combination of
hurricanes and earthquakes (Bangladesh, Nepal, Mozambique, Honduras, El Salvador),
drought (Burkina, Ethiopia, Mali), or a combination of hurricanes, droughts, and floods
(Kenya, Nicaragua).
3
Table 1: Selection of the projects included in the sample, per country and per continent
Natural Integrated
Agricultural Water resource rural
research management management development
(AR) (WM) (NRM) (IRD) Total HDI rank
Pakistan 1 2 2 5 138
India - 2 1 1 4 124
Bangladesh 4 4 145
Nepal 2 2 142
Philippines 1 1 77
Asia 1 6 6 3 16 (35%)
Burkina 4 4 169
Ethiopia 4 4 168
Mali 2 2 167
Kenya 1 1 2 134
Egypt 2 2 115
Benin 1 1 158
Cape Verde 1 1 100
Mozambique 1 1 170
Africa 5 2 1 9 17 (37%)
Bolivia 2 1 2 5 114
Nicaragua 2 2 4 118
Honduras 1 1 116
Costa Rica 1 1 43
El Salvador 1 1 104
Ecuador 1 1 93
L. America 2 1 4 6 13 (28%)
8 (17%) 9 (19%) 11 (24%) 18 (40%) 46 (100%)
Source: Database containing information from 46 evaluation reports (see Appendix), and Human
Development Report 2004.
Of the projects under review, most of the AR and IRD projects were carried out in
Africa (for example, Mali, Kenya, Benin, Mozambique, Burkina Faso, Ethiopia). WM
projects were mainly concentrated in Asia (for example, Bangladesh, India), while the
NRM projects, although more dispersed, were mostly found in Central America (for
example, Nicaragua, Costa Rica, El Salvador) and Asia (for example, Pakistan, Nepal,
India, Philippines). Even though it is difficult to make a strict distinction between the
different categories,4 the goals and scope of activities (described later in Table 2) can be
summarized as follows.
4 In making a classification of the projects, it is important to realize that there is a lot of overlap of goals.
Seven projects fall between IRD and NRM projects; four projects classified as WM have the
characteristics of IRD projects; and five projects classified as IRD were mainly focused on institutional
development (which is not mentioned as one of the administrative categories). Project classifications are
often arbitrary.
4
3 Agricultural research projects
Most of the AR projects under review were initiated in an attempt to fill the vacuum in
agricultural research that was created during the neoliberal period when many of the
national research institutes and the extension services were closed down, and no further
investments were made in agricultural development. Many are national projects with
dispersed intervention areas, and the major concern is to deal with the problem of low
agricultural returns.
In many projects, attempts were made to introduce farming-system research (to replace
the ‘old-fashioned’ crop-related research), and much emphasis was put on developing
new packages more suitable for the poor and with more concern for environmental
factors. Initially, many AR projects paid little attention to economic aspects and often
had a top-down orientation. In the course of time, along with the tendency to give more
priority to extension (rather than research), more attention was paid to linking up with
the target group and introducing a more participatory approach. Institutional linkages
were rather weak, even though the gap between policy making and research was often
seen as a bottleneck.
Water management has been on the DGIS agenda for some time, and was often
presented as a necessary ingredient for agricultural modernization. Interestingly,
whereas many donors (and the DGIS) had refrained from intervening in politically
sensitive issues, such as land reform, large sums of money have been invested in the
construction of irrigation systems, the feasibility of which was expressed in cost-benefit
relations. Little mention was made of the political dimensions of the water project,
namely the distribution of water.
5
5 Integrated rural development projects
Integrated rural development projects were very popular in the 1980s as the best
instrument for alleviating the situation of the rural poor in the most marginalized areas
of the developing world (Livingstone 1979; Zoomers and Geurten 1991). Growing
disillusionment with technocratic and bureaucratic approaches to rural development
(green revolution, agricultural colonization, land reform, etc.) resulted in the conclusion
that it was not appropriate to attack single constraints through top-down planning and
narrow sector programmes. It was recognized that rural development comprises the
interaction of a large number of interrelated activities, and most IRD projects are
relatively broad. Attention is paid to agriculture, but also to roads, irrigation, schooling,
sanitation, credit and/or small-scale industrialization. ‘IRD involves all the things that
can most improve the living conditions of the rural masses’ (Gebregziabher 1975, cited
in Zoomers and Geurten 1991: 195).
The IRD projects under review were often carried out in marginal and isolated regions,
concentrated in dry desert areas with problems of famine and ecological degradation.
Projects aim at supporting subsistence farmers in guaranteeing food security, usually in
combination with a wide range of other activities. Most IRD projects use a bottom-up
approach, aimed at simultaneous development of different activities in accordance with
local needs and circumstances. Many IRD projects developed in isolation from national
policy (IRD being too broad to be covered by a single ministry).
Many of the NRM projects were introduced in the early 1990s, in response to a call by
the United Nations Conference on Environment and Development (UNCED) to pay
more attention to environmental sustainability. Although environmental problems
(erosion, overgrazing, climate change, etc.) affect large parts of the developing world,
all NRM projects reviewed have a direct link with forest resources, and were located in
subtropical rainforests, often in old colonization areas.5 Whereas earlier DGIS
interventions had been aimed at helping governments to carry out colonization projects
(settlement of colonists), most NRM projects focused on reducing deforestation and
environmental degradation, while helping small colonist farmers to improve forest
5 In the 1970s and 1980s, governments, and especially Latin American governments, often supported by
international donor organizations, carried out a policy of agricultural colonization policy. The expansion
of the crop farming area was seen as an appropriate strategy for economic growth; in addition it would
provide landless workers with new land, thus solving the social problems in the countryside. Some years
after their initial cultivation, however, many colonization areas started to show severe problems: Forest
areas were not very suitable for permanent cultivation, and in many of these zones, land conflict appeared
between different groups (cattle herders, agro-industries, indigenous groups, colonists, lumber enterprises,
etc.).
6
management.6 Many of the NRM projects underwent changes during their lifetime; the
focus shifted from tree planting to forest management. Later, many of these projects
broadened their scope, and paid more attention to the economic dimensions and/or
social aspects (including gender) while introducing a more participatory approach; there
was some shift from nature to economic activities and people. Because of the
institutional weakness in the field of natural resources (no specialized ministries), many
of these projects were rather isolated in the beginning; but in the course of time—along
with decentralization policies in the countries involved—the projects developed closer
relationships with the local government.
Most of the projects have relatively small areas of interventions: 38 projects only
covered one single municipality, a district, or a watershed (with a limited target
population); only eight projects (AR) were national projects with scattered intervention
areas. In our sample, but also in other cases, there are no examples of projects working
together within the same project area.8 DGIS-funded projects are in general mostly
dispersed, and there is little or no geographic clustering.
6 It is striking that relatively little attention is given by DGIS to environmental problems in other types of
areas, such as deserts, mountains, etc. NRM interventions show a strong bias in favour of forest areas.
7 Initially, IRD projects were broad but they became more focused; NRM projects started with an
environmental focus, but the scope widened in the course of time.
8 All the projects were aimed at different regions (i.e., no overlap in area of intervention between WM-
NRM-AR-IRD projects. All seem to have had their own target area, with the exception of the Malakan
Division (Pakistan) and the Bluefield region (Nicaragua).
7
Table 2: Characterization of the projects
Field of activities Dominant goal Type of problem Type of target group Location (also Table 1)
40% (N=18) IRD-related range of Disaster prevention and Famine and Subsistence farmers, Often in dry-desert areas
IRD activities food security marginalization marginalized, support of sub-Saharan Africa
related to food crops
(> 5 activities) with Livelihood improvement Ecological degradation
tendency to become more
restricted Institutional strengthening
24% (N=11) Range of activities related Environmental goals, Deforestation and Crop producers, Mainly in old colonization
NRM to agro-forestry/social combined with livelihood ‘cattleization’ diversified, often in areas; forest frontiers of
forestry improvement, settlement, combination with Central America
(2-5 activities) with land distribution Environmental problems migration, support related
tendency to become to agro-forestry
broader
19% (N=9) Watershed management, Agricultural production, Water logging and Crop producers, diversified Mainly in Asian
WM irrigation and flood control, livelihood improvement, salination and specialized, support watersheds
often with infrastructure settlement, land related to WM and
(< 2 activities), relatively development distribution Flooding and groundwater irrigation.
specialized problems
Institutional strengthening
17% (N=8) Farming systems research Agricultural production Low agricultural returns Crop producers. Dispersed, often national
AR and crop development, Diversified, crop-specific projects
often related to dry land Institutional strengthening support, related to cotton,
(< 2 activities), relatively farming soya, potatoes, rice,
specialized
Source: Database containing information from 46 evaluation reports (see Appendix).
8
There are considerable differences in geographical characteristics of the project areas
(Table 3): 22 per cent of the projects were carried out in the savannahs of the Sahel, and
11 per cent in the high mountains of the Andes; 37 per cent of the projects focused on
deforested areas in the tropics/subtropics, while 22 per cent were carried out in
watersheds and/or on flood plains; in 8 per cent of the cases the project area was not
specified. In the project documents little information is usually available about the
characteristics of the target group: 23 per cent of the projects focused on subsistence
farmers; the rest speak about ‘smallholders’ (that is, a very diverse group of people). It
is striking, however, that little or no attention is given to other groups as fishers, cattle
herders, landless people, or non-sedentary groups (nomads), even though these often
form a considerable part of the rural population, also within the project areas.
A noticeable characteristic of the projects under review is their long duration. The
average duration of the projects under review was 10.7 years (range: 3-25 years); almost
50 per cent lasted between 8 and 11 years which was often not foreseen in the original
design. Many of the projects experienced a chain of extensions. The project team stayed
longer than planned, in response to disappointing results or to problems related to
transferring the results to counterpart organizations. In spite of being extended several
times, 52 per cent of the projects ended without a clear exit strategy; 20 per cent ended
‘unexpectedly’ as a direct consequence of changing donor policy (including projects
that were stopped with the introduction of the SWAp policy).
During the lifetime of the projects, important transformations took place in 80 per cent
of the cases. These transformations concerned the aims of the project (65 per cent of
cases, for example shifting from environmental to social goals); the approach (50 per
cent; e.g., with more participation, more attention to gender relations), the delimitation
of the project area (35 per cent; often a reduction for reasons of efficiency); changes in
counterpart organizations (52 per cent), and/or in donor structure and funding situation
9
(13 per cent). Most of the DGIS projects were carried out bilaterally, without the
involvement of other donors.9 Almost all DGIS projects were traditional in the sense
that 83 per cent were directed by expatriates, usually under the guidance of a specialized
Dutch consultancy firm or development organization. Salary costs (project staff and
missions) form the bulk of the budget, although there were considerable differences
between the different projects. In WM, for example, large sums were spent on
infrastructure, whereas in IRD much was also spent on credit programmes, sanitation,
and/or education. There was considerable diversity in the project budgets.10
As regards the extent to which the projects worked in isolation and/or had links with the
government and/or non-governmental organizations (NGOs), there is a clear difference
between AR and WM projects on the one hand (direct contact with line ministries,
central level) and IRD and NRM projects on the other (networking mainly with NGOs
and community organizations, local level). The ‘vertical integration’ of most projects
was very weak. It is only during the final period (along with the introduction of the
SWAps) that projects became incorporated into the policy structure—projects started to
link up with ministries, and this had positive implications for their capacity to influence
mainstream policy.
In order to make a distinction between the projects on the basis of failure or success, we
first made an assessment on the basis of the DAC evaluation criteria (see Box 1) and the
GAVIM goals (see Box 2) both adopted by DGIS as guidelines for field evaluations in
the early 1990s. The DAC criteria give an impression of the relevance, effectiveness,
efficiency, impact, and sustainability of the individual project results (OECD/DAC
1991). The GAVIM criteria help to assess whether the projects gave sufficient priority
to dimensions such as poverty, environment, gender, governance, and institutional
9 Exceptions are projects where co-financing took place with the UN World Food Organization, UNDP,
International Fund for Agricultural Development, EU, and the World Bank.
10 NRM and AR projects were relatively low cost, as opposed to the IRD and WM projects.
10
BOX 1: DAC CRITERIA FOR PROJECT EVALUATION
Relevance: The extent to which the aid activity is suited to the priorities and policies of the target group,
recipient, and donor: Are the objectives valid? Are the activities and outputs consistent with the overall
goal and with the attainment of the objectives? Are the activities and outputs of the program consistent
with the intended impacts and effects?
Effectiveness: A measure of the extent to which an aid activity attains its objectives: Were the objectives
achieved? What were the major factors influencing achievement/non-achievement?
Efficiency: Measures the outputs, both qualitative and quantitative, in relation to the inputs: Were the
activities cost-efficient? Were the objectives achieved on time? Was the project implemented in the most
efficient way compared to the alternatives?
Impact: The positive and negative changes produced by a development intervention, directly or indirectly,
intended or unintended: What has happened as a result of the project? What real differences have the
activities made to the beneficiaries? How many people have been affected?
Sustainability: This is concerned with measuring whether the benefits of an activity are likely to continue
after donor funding has been withdrawn. Projects need to be environmentally as well as financially
sustainable: What happened to the benefits of the project after funding ceased? What were the major
factors that influence the achievement or non-achievement of sustainability of the program or project?
Source: OECD-DAC (1991).
11 The rating process, presented here as very straightforward, was in practice very complex and rather
subjective. We took the ratings given by the different teams, but had to make adaptations based on
differences in interpretation (each team having their own reference framework) but also the lack of
consistency between the teams. By making an assessment of the ratings in comparison with the contents
of the reports, the author tried to eliminate subjectivities and make the information more comparable. The
11
success, namely it was relevant, effective, efficient, had a good impact and led to
sustainable results. A low GAVIM rating indicates that sufficient priority was given to
goals such as poverty alleviation, environment, gender, and so forth. Higher DAC
scores indicate a negative performance, and a high GAVIM rating shows that
insufficient attention was paid to the GAVIM dimensions.
Table 4 shows that the best DAC results were achieved in terms of relevance and
effectiveness (1.73 and 2.24, respectively); the results for sustainability, impact, and
efficiency are less favourable (2.73, 2.51, and 2.38). With respect to the GAVIM goals
(Table 5), the scores for environment, gender, and poverty are relatively favourable
(2.14, 2.17, and 2.19), while those for institutional development and governance are less
favourable (2.40, 2.55).
ratings give some indication of the performance of the projects and is used as a starting point of further
analysis.
12
The best GAVIM results were achieved by the IRD projects (in spite of relatively
negative scores for governance and institutional development); this positive picture is
supported by a rather positive DAC score for relevance (but negative scores for
effectiveness, impact, and sustainability). AR projects have relatively positive results
for DAC (relevance and effectiveness), in spite of showing a rather negative score for
sustainability; these projects also show rather positive GAVIM scores for institutional
development and governance (but negative for gender, environment, and poverty).
WM projects have scores for environment (GAVIM) and relevance (DAC), but negative
GAVIM scores for governance and gender. The NRM projects were the least successful
projects: They have negative scores for efficiency, effectiveness, impact, and
sustainability (DAC), and negative ratings for all GAVIM goals (Tables 4 and 5). As for
the different regions, Tables 6 and 7 show that the projects in Africa achieved relatively
favourable GAVIM scores (poverty and environment), and DAC scores (relevance). At
the other extreme, projects in Central and South America have negative GAVIM scores
for governance and institutional development, as well as negative DAC scores for
impact and sustainability (in combination with a relatively favourable score for
relevance).
13
7.2 Successes versus failures
For the purpose of this study, we divided the projects into two groups: namely the
successful projects (i.e., projects with DAC and GAVIM scores of less than 11 and 8,
respectively) and the unsuccessful projects (projects with DAC and GAVIM scores
higher than or equal to 11 and 8) (see Table 8). These cut-off points were chosen
because they allowed us to make a selection of the top 20-30 per cent of the most and
least successful projects (reflected by consistent patterns of extreme negative and
positive scores, respectively). Projects with mixed results were not taken into
consideration.12
The most successful projects (a group of 10) have positive DAC scores for relevance
(1.45), efficiency (1.85), and effectiveness (1.90), and favourable GAVIM scores for
environment (1.67), gender (1.75), and poverty (1.80). At the other extreme, the failures
(16 projects) have negative DAC scores for sustainability (3.16), impact (2.90),
efficiency (2.90), and effectiveness (2.63), and unfavourable GAVIM scores for
governance (2.81) and institutional development (2.70). As for the characteristics of the
more successful projects, there is some concentration of IRD projects carried out in
relatively marginal and isolated areas in poor countries (three in Ethiopia, one each in
Burkina, Honduras, and India): Successful projects were relatively strong in helping
subsistence farmers to improve their food security situation, in contributing to disaster
prevention, and in working with the poorer groups of subsistence farmers; other
successful projects were in Pakistan (2 x NRM), and in India and Bangladesh (2 x
WM).
At the other extreme (the failures), there is some over-representation of NRM projects
carried out in the colonization areas of Central America (three in Nicaragua, and one in
Costa Rica) that were aimed at solving the problems of deforestation and/or overgrazing
(introduction of cattle); most involved working with farmers with diversified livelihoods
(farming combined with migration, partly as an exit strategy). Other failed NRM
projects were in Nepal, India, and the Philippines. Also the IRD projects carried out in
Nicaragua, Ecuador, Bolivia (2 x), and Burkina Faso were not a success; other failures
were found in India and Bangladesh (3 x WM), and in Benin and Pakistan (2 x AR).
The best results were thus found in the most deprived regions in Africa; this is mainly
12 We only looked at the projects showing extreme scores; and did not take into account the projects with
‘opposite’ DAC and GAVIM scores. This decision was made to reduce subjectivities and inconsistencies
(see footnote 11). We decided to select 26 cases with the clearest and most consistent results, and used
this as an input for further analysis.
14
Table 8: Ten most successful and 16 least successful projects on the basis of mean DAC-GAVIM scores
Most successful projects Least successful projects
Name of project Country Type of project DAC+ GAVIM Name of project Country Type of project DAC+GAVIM
ERP Pakistan NRM-institutional 16 SARC-TSARDD Philippines NRM-IRD 22.5
development
PDI/Z -PDL/Z Burkina Faso IRD 17 CHOROTEGA Costa Rica NRM-IRD 22.5
BAREAP Ethiopia IRD 16 PRODES Nicaragua IRD-NRM 24.5
GIRPDP Ethiopia IRD 16.5 PROCODEFOR Nicaragua NRM-IRD 24
LEMPIRA Honduras IRD 17.5 PSB/PB Burkina Faso IRD 22
ADTDP India IRD 11 Chuquisaca Centro Bolivia IRD-NRM 20
SUPAK Ethiopia IRD-institutional 16 Fortalecimiento org. Bolivia IRD-Inst. Dev. 23.5
development
MALAKAND Pakistan NRM 18 Servicio postcosecha Ecuador IRD-Inst. Dev 26
CDSP Bangladesh WM/IRD 19 NRAP-ISWASRI Pakistan AR 23.5
BENGAL TERAI India WM/IRD 18 Bundelkhand India NRM/WM 27
Mechi Hill Nepal NRM-IRD 26
PIE Monte Nicaragua NRM 29
RAMR-PARP Benin AR 21.5
SRP Bangladesh WM 21.5
CPP Bangladesh WM 21.5
HOPP India WM 25
Source: Database containing information from 46 evaluation reports (see Appendix).
15
reflected by the higher scores for poverty (GAVIM) and relevance (DAC). The most
negative results were in the colonization projects in Central and South America.
Even though DAC and GAVIM criteria help to get a global picture of projects’
successes versus failures, there are several limitations in the use of such indicators. A
negative evaluation in terms of DAC or GAVIM will not necessarily mean that the
original project goals were not achieved: In many cases, DAC or GAVIM criteria are
used ex-post facto for purposes of evaluation, while in the original project document no
mention was made of aspects such as gender or environment. Also the opposite is true; a
positive evaluation in terms of DAC and GAVIM will not necessary mean that the
project was in line with the livelihood priorities of the population (or were appreciated
by the population). Projects aimed at improving farm income might be called a success
in terms of DAC criteria or GAVIM goals, but this view is not necessarily shared by the
majority of the population who might have had a preference for other, non-agricultural,
activities (Reardon et al. 2001). There might have been better, more effective solutions
or a shorter road to poverty alleviation. In many cases, the scores are not so much a
reflection of the performances of the project, but much more a reflection of the
complexity of context: GAVIM and DAC criteria tend to evidence more favourable
results in settings that are subject to climatic disaster or political upheaval, given that
the scores for poverty (GAVIM) or relevance (DAC) will show more striking, positive
results. At the same time, however, improvements in the local situation will not
necessarily be a proof of the ‘success’ of the project; improvements may be nothing to
do with the project, but with political peace or stability, or recovery after a war or a
natural disaster. In addition, it is myopic to make an assessment of the success or failure
of projects without taking into account cost-benefit aspects of project implementation.
Although natural resource management projects might have generated less positive
results than other projects, how much money was spent, and who benefited? Low-cost
failures cannot be automatically compared with high-cost failures. It is important to take
such dimensions into account when referring to DAC and GAVIM scores in
evaluations.
An important aim of this study was to make an aggregate-level assessment of the factors
contributing to the success or failure of a project. To do so, an investigation was made
into what patterns could be identified regarding the factors affecting the success or
failure of activities, and the extent to which these factors are region- or theme-specific.
There are a number of interesting correlations between DAC and GAVIM scores and
other variables available in the data material that give an indication of the kind of
factors that explain the success or failure of project interventions.
First, there is correlation between DAC and GAVIM scores, even though they measure
different things (the correlation between the total DAC and GAVIM score amounts to
16
572**). There is a strong correlation between the effectiveness of projects (DAC) and
the GAVIM score for governance (710**); there is also correlation between the total
DAC score with GAVIM scores for governance (626**), institutional development
(505**), and poverty (427**); there is less correlation with GAVIM scores for
environment (351*) and gender (no correlation). Other correlations were found for the
total GAVIM score and the DAC scores for efficiency (550**) and effectiveness
(403**), impact (395*), and sustainability (376*); no correlation was found with the
DAC score for relevance. In spite of measuring different dimensions, DAC and GAVIM
scores often point in similar directions and seem to be mutually linked.
Second, there are significant correlations between the GAVIM and/or DAC scores and a
number of project characteristics; Table 9. The most relevant projects are well focused
(422**) and sufficiently large (DGIS-contribution; -398**). The impact of a project
seems to depend mainly on the degree of vertical integration and/or the availability of
micro-macro linkages (406**), as well as on the level of investments (not too large;
349*) and the degree of flexibility (allowing for timely adaptations without losing their
focus; 310*), and are relatively small and inexpensive (DGIS-contribution; 349*). The
effectiveness of projects are mainly related to the context (HDI score; -310*). Finally,
efficiency is linked with the availability of systems for monitoring and evaluation (-393*
and -386*, respectively).
17
The achievement of GAVIM goals seem to be related to the horizontal
linkages/networking capacity (-413**, -370*, -365*), and a sufficient degree of
stability-flexibility (585**), as well as to the project focus (not too specialized; -322*).
Also the degree of vertical integration (micro-macro; 508*) and the context (HDI: -339*
and 336*) are important for the GAVIM scores. Projects active in networking show a
good performance on gender and environment. Projects carried out in complex
situations show positive results for institutional development. Stable and well focused
18
projects with a sufficient degree of vertically integration show positive scores for
governance. And broader projects (no specialization) have the best poverty scores. The
correlation between the HDI rank and the DAC and GAVIM scores confirms the earlier
conclusion that in more problematic situations (with lower levels of human
development according to the HDI), the DAC and GAVIM scores are more positive
than in countries with lower rankings. This picture of patterns of factors explaining the
success or failure of projects is confirmed by a more detailed analysis of the critical
success and failure factors mentioned in the different evaluation reports. The following
list reflects the kind of factors identified; Table 10.
• Target group orientation (mentioned 60 times in the evaluation reports).
Successful projects are said to have a homogeneous and clearly defined target
group, with sufficient attention paid to empowerment and tangible benefits.
• Institutional characteristics and organizational setup (mentioned 56 times).
Successful projects are locally embedded: They involve all the crucial actors and
counterparts with a sufficient level of involvement of the grassroots level (bottom
up). Staff quality and continuity are also very important.
• The policy context (mentioned 54 times). In successful projects often there is
consistency between project goals and national policy with a minimum of donor-
driven changes
• The sociopolitical circumstances (mentioned 14 times), economic environment/
infrastructure (23 times), and ecology (16 times). Project results are often
negatively influenced by the unfavourable macroeconomic conditions and pricing,
and by a poor ecological situation.
• Project design, planning, and implementation (mentioned 52 times) play an
important role; successful projects show consistency between goal and activities,
and have appropriate technologies. It is also important to have a process approach
with a sufficient degree of flexibility.
• The human resources and project team organization (mentioned 22 times).
Successful projects have a professional staff with sufficient continuity (staff
turnover is often a problem).
More specifically, the most important determinants of success or of failure are the
consistency between project goals and national policy (E, mentioned 35 times; 19 times
as a reason for failure; 16 times as a reason for successes), the conceptual setup and/or
lack of consistency (B, mentioned 27 times; 21: failure, 6: success), the local
embeddedness/involvement of the crucial actors or counterparts (D, mentioned 26
times; 16: failure, 10: success), target group orientation (C, mentioned 25 times; 16:
success, 9: failure), and empowerment activities (C, mentioned 22 times; 12: failure, 10:
success).
19
More specifically, success was mainly attributed to: Having a clear target group focus
and paying sufficient attention to empowerment (C); local ownership and the
involvement of crucial actors/counterparts (D); consistency between project goals and
national policy (E); and having a process approach/sufficient flexibility in project
implementation (B). The failure of projects was mainly explained in terms of
inappropriate project design and lack of consistency (B), an unfavourable policy
environment (E), and/or other unfavourable external circumstances (F, H, G) and lack
of continuity of staff/high staff turnover (A).
The above indicates that the DGIS projects have changed considerably over the last
period, and that many projects are wrongly described as static and fragmented. Many
projects show rather positive results, even though performance depends very much on
the local circumstances (target group, available actors etc.). The performance of projects
will very much depend on the project design, the institutional landscape, human
capacities, etc. At the same time, however, the move from projects to SWAps seems a
step in the right direction. This new policy context will offer better opportunities to
apply more co-ordinated, multisectoral approaches, with more ownership and a better
degree of vertical integration. In some cases change is the result of DGIS reacting to
external changes and in other cases is a result of lessons learned. In many cases there is
not a strict separation between the project and programme approach. Many projects
during their lifetime moved into the direction of a SWAp (became more focused,
adopted a process approach, were active in networking, strengthened ‘local’ ownership
etc.).
In the current debate about how to improve the performance of development co-
operation, much emphasis is given to the subject of organizational learning—the lack of
solid knowledge about the impact of aid’ is nowadays mentioned as problematic, having
a negative impact for the effectiveness of aid (Grinspun 2001). According to Carlsson
and Wohlgemuth (2000: 7), learning in development co-operation is difficult due to five
factors that seem to be particularly prominent: political constraints; the unequal nature
of aid relationship; problems internal to the organization of the aid agency; organization
and capacity of the recipient; and sources of knowledge and quality of information. In
order to improve the learning experiences of the stakeholders active in development co-
operation (‘do we learn from our experiences and do we feed that knowledge back into
improved practices?’), much will depend on the availability of data collections
systems.13 Much priority is given to social analysis: improving feedback and
13 Only 13 per cent of the reviewed projects had some kind monitoring system. To the extent that
information is available about the DGIS projects, this mainly consist of simple output variables (e.g.,
number of people trained, improvements realized, settlements helped, hectares under irrigation, trees
planted, roads improved, etc.), but there was no systematic collection of information about the impact
20
communication practices to promote an evaluation culture and to implement country
programmes and joint evaluations; and to promote partnerships in evaluations, design
and implement performance measurement systems (OECD-DAC 1998:2; Bamberger
and Hewitt 1986).
Even though these kinds of transformations will be important inputs for learning, we
would like to raise a number of ‘hidden’ issues that are often neglected in the current
debate, and which need to be solved in order to improve the performance of
development co-operation.
9 A strategic mission?
Looking at the goals and activities of the various projects, it is striking that project aims
are usually very vague, and it is not so clear why priority was given to a particular
strategy (instead of alternative ways to achieve poverty alleviation). In addition, during
the final decade the consistency between goals and activities even deteriorated, amongst
others due to the multiplication of evaluation criteria (for example GAVIM, etc.). Many
projects in the 1990s moved away from their original project goals, without adjusting
the strategy how they wanted to contribute to poverty alleviation. Projects which,
according to their mission statement, were mainly aimed at ‘helping farmers to improve
their income situation’ focus on activities such as tree planting or institutional
development even though this will probably not be the shortest way to poverty
alleviation. Projects aimed at supporting the rural poor to escape from poverty (still)
21
focus on introducing new crop varieties or improving irrigation as a strategy for
agricultural development. There are no interventions related to migration even though
this would have helped people to accumulate capital in a much easier way. There was
often no systematic search for the best solution given the whole range of opportunities
and the aspirations of the population.
In project documents, little attention is usually given to the context, or more specifically
the reality that most of the projects concentrate in areas where floods, hurricanes, etc.,
are part of normal life (people will have to deal with such situations, even though they
are calculated as a risk). The projects under review were carried out in post-conflict
regions, or in areas hit by natural disasters, but in almost all the project documents
targets are set on the basis of the ‘lucky’ situation (no disasters as the 0-situation). In
evaluations, the war situation, conflicts, or the incidence of drought or flood, are
mentioned as reasons for not achieving the expected results, which is often the reason
for prolongation. In areas of development co-operation work, situations of war, drought
and hurricanes are insufficiently seen as ‘normalities’; project goals are formulated in
terms of poverty alleviation, even in circumstances where priority should be given to
earlier goals (such as political stabilization). Given the complexity of the project
situation, it is striking that so little attention is paid in most projects to disaster
prevention. Every time a new hurricane occurs donor funds are spent on reconstruction
as if it were mere bad luck and not as part of the normal life situation.14 Many projects
have negative results as a direct consequence of setting unrealistic goals. It is important
to mention the problematic context in a more explicit way, and to see how interventions
could help people to better anticipate future emergencies (preventing problems, instead
of solving them).
10 Final reflections
14 Only in the case of WM (flood control) and IRD (food security) is there some reflection on how the
project could help people to better cope with such misfortunes.
22
attention should be given to how to make development co-operation more strategic and
strengthening the link with the local context.
15 ‘In the ideal learning situation there is a substantial amount of information created on both sides. This
information is often scattered among different sources, which does not make it easy to access it and obtain
a good overview. The information needs to be aggregated and synthesized in order to provide a
comprehensible and available picture of the current situation’ (Carlsson and Wohlgemuth 2000: 9).
23
Appendix
Agricultural research
24
Water management
25
References
Bamberger, M., and E. Hewitt (1986). ‘Monitoring and Evaluating Urban Development
Programmes. A Handbook for Program Managers and Researchers’, World Bank
Technical Papers 53, World Bank: Washington DC.
van Dijk, J.W.M., D.F. Bryceson, P. Howard, J. Oorthuizen, and A. Zoomers (2005).
‘Rural Development Project Performance: A Review of 46 Evaluation Studies of
DGIS-funded Projects in the Themes Agricultural Research, Natural Resource
Management, Water Management and Area Development and the Implications for
the Sector Wide Approach’, Consultancy Report for DGIS DDE/NB on behalf of
CERES Research School for Resource Studies for Development: Utrecht.
Carlsson J., and L. Wohlgemuth (eds) (2000). Learning in Development Co-operation,
Almqvist & Wiksell International: Stockholm.
DGIS (2000). ‘Working Document GAVIM and the Sectoral Approach’, mimeo
(www.minbuza.nl) Ministry of Foreign Affairs: The Hague.
DGIS (2003a). ‘Aan elkaar verplicht. Ontwikkelingssamenwerking op weg naar 2015’
(www.minbuza.nl), Ministerie voor Buitenlandse Zaken (DGIS): Den Haag.
DGIS (2003b). ‘Sectorale Benadering, Organiserend Principe Voor De Bilaterale
Ontwikkelingssamenwerking’, Groeidocument 2, Ministerie voor Buitenlandse
Zaken (DGIS), Steungroep Sectorale Benadering: Den Haag,
26
OECD-DAC (1991). DAC Principles for Evaluation of Development Assistance,
OECD-Development Assistance Committee: Paris.
OECD-DAC (1996). Shaping The 21st Century: The Contribution Of Development
Co-operation, OECD-Development Assistance Committee: Paris.
OECD-DAC (1998). Review of the DAC Principles for Evaluation of Development
Assistance, OECD-Development Assistance Committee: Paris.
Reardon, T., J. Berdegué, and G. Escobar (2001). ‘Rural Non-farm Employment and
Incomes in Latin America: Overview and Policy Implications’, World Development
29(3): 395-409.
Schulpen L. (ed.) (2001). Hulp In Ontwikkeling: Bouwstenen Voor De Toekomst Van
Internationale Samenwerking, van Gorcum: Assen.
27